• 한국정보통신학회논문지
• /
• 제17권2호
• /
• pp.453-460
• /
• 2013

본 연구에서는 클라우드 환경에서 기하급수적으로 증가하는 전사적 정보 연계를 위한 시맨틱 기반 개념 망을 이용하여 전사적 데이터들의 효율적 연계와 활용이 가능하도록 프레임워크를 제안한다. 개념 망은 기존 온톨로지에 접근하는 방식은 유사하지만, 사용자가 보다 효율적으로 정보 연계 검색을 하고자 객체와 개념 사이의 연관성을 구축 한다. 본 논문에서는 개념 망을 3가지로 구분하여 제안 프레임워크에 적용한다. 본 연구의 개념 망은 마스터 정보 개념 망, 키워드 개념 망, 그리고 비즈니스 프로세스 개념 망을 기반으로 온톨로지 형태로 구축된다. 이 개념 망은 사용자 요구사항에 따라 데이터들 간의 연관성을 기준으로 하여 검색 및 활용을 가능하게 한다. 그리고 마스터 정보 개념과 키워드 개념이 결합되어 검색 키워드의 빈도 및 카테고리의 빈도 추적을 제공함으로써, 사용자의 검색의 편의성과 신속성을 향상시킬 수 있도록 하였다.

• Nuclear Engineering and Technology
• /
• 제50권3호
• /
• pp.327-339
• /
• 2018

This article presents a new global-local hybrid coarse-mesh finite difference (HCMFD) method for efficient parallel calculation of pin-by-pin heterogeneous core analysis. In the HCMFD method, the one-node coarse-mesh finite difference (CMFD) scheme is combined with a nodal expansion method (NEM)-based two-node CMFD method in a nonlinear way. In the global-local HCMFD algorithm, the global problem is a coarse-mesh eigenvalue problem, whereas the local problems are fixed source problems with boundary conditions of incoming partial current, and they can be solved in parallel. The global problem is formulated by one-node CMFD, in which two correction factors on an interface are introduced to preserve both the surface-average flux and the net current. Meanwhile, for accurate and efficient pin-wise core analysis, the local problem is solved by the conventional NEM-based two-node CMFD method. We investigated the numerical characteristics of the HCMFD method for a few benchmark problems and compared them with the conventional two-node NEM-based CMFD algorithm. In this study, the HCMFD algorithm was also parallelized with the OpenMP parallel interface, and its numerical performances were evaluated for several benchmarks.

• 국제초고층학회논문집
• /
• 제9권4호
• /
• pp.351-360
• /
• 2020

Monitoring civil structures periodically is necessary for ensuring the fitness of the structures. Cracks on inner and outer surfaces of the building plays a vital role in indicating the health of the building. Conventionally, human visual inspection techniques were carried up to human reachable altitudes. Monitoring of high rise infrastructures cannot be done using this primitive method. Also, there is a necessity for more accurate prediction of cracks on building surfaces for ensuring the health and safety of the building. The proposed research focused on developing an efficient crack classification model using Transfer Learning enabled EfficientNet (TL-EN) architecture. Though many other pre-trained models were available for crack classification, they rely on more number of training parameters for better accuracy. The TL-EN model attained an accuracy of 0.99 with less number of parameters on large dataset. A bench marked METU dataset with 40000 images were used to test and validate the proposed model. The surfaces of high rise buildings were investigated using vision enabled Unmanned Arial Vehicles (UAV). These UAV is fabricated with TL-EN model schema for capturing and analyzing the real time streaming video of building surfaces.

• 드라이브 ㆍ 컨트롤
• /
• 제20권4호
• /
• pp.92-99
• /
• 2023

Autonomous vehicles should recognize and respond to the specified speed to drive in compliance with regulations. To recognize the specified speed, the most representative method is to read the numbers of the signs by recognizing the speed signs in the front camera image. This study proposes a method that utilizes YOLO-Labeling-Labeling-EfficientNet. The sign box is first recognized with YOLO, and the numeric digit is extracted according to the pixel value from the recognized box through two labeling stages. After that, the number of each digit is recognized using EfficientNet (CNN) learned with the virtual environment dataset produced directly. In addition, we estimated the depth of information from the height value of the recognized sign through regression analysis. We verified the proposed algorithm using the virtual racing environment and GTSRB, and proved its real-time performance and efficient recognition performance.

• 한국전자통신학회논문지
• /
• 제19권3호
• /
• pp.577-582
• /
• 2024

국내 펫 보험은 2007년 첫 도입되어 현재 2024년 지금까지 많은 보험상품들이 생겼고 펫 보험 시장은 매년 증가하고 있는 추세이다. 하지만 실상은 2022년 기준 펫 보험 가입률은 전체 반려인의 0.8%이며 반려인들은 비싼 보험료 및 보장내역, 까다로운 가입 기준으로 인해 펫 보험 가입을 꺼리고 있다. 본 논문에서는 반려동물 안구질환 및 질환의 위치를 인식하고 건강나이를 예측 가능한 모델링을 제안한다. 먼저 EfficientNet을 활용해 반려동물의 안구질환을 인식하고 OpenCV를 활용 질환의 발병 위치와 크기를 인식하여 반려동물의 건강나이를 산출한다. 산출된 해당 건강나이를 바탕으로 보험사에서 펫 보험료 산정 시 보조하는 역할을 하고자 한다. 이 모델링은 반려동물 안구질환 및 건강나이로 합리적인 펫 보험 가격 산정 보조가 가능하다.

• 한국어류학회지
• /
• 제20권3호
• /
• pp.228-232
• /
• 2008

To investigate a proper sampling method for anchovy eggs and larvae in coastal waters of the South Sea, replicated samplings were made by different towing methods with different sampling gears and compared in terms of abundance and length composition. There was no significant difference in abundance in samples from vertical and oblique tows with a ring net. The abundance by replicated vertical tows with a ring net was not significantly different, but significant difference in abundance among sampling stations were found. The ring net sampled anchovy eggs in significantly greater numbers than collected by a NORPAC net, but both gears were not effective in obtaining quantitative samples of anchovy larvae larger than 3 mm. Therefore, samples by vertical tows with a ring net during the day at various stations is more efficient at estimating the density of anchovy eggs in an area compared to replicated sampling at a single station.

• ETRI Journal
• /
• 제41권6호
• /
• pp.771-781
• /
• 2019

We propose the VS³-NET model to solve the task of question answering questions with machine-reading comprehension that searches for an appropriate answer in a given context. VS³-NET is a model that trains latent variables for each question using variational inferences based on a model of a simple recurrent unit-based sentences and self-matching networks. The types of questions vary, and the answers depend on the type of question. To perform efficient inference and learning, we introduce neural question-type models to approximate the prior and posterior distributions of the latent variables, and we use these approximated distributions to optimize a reparameterized variational lower bound. The context given in machine-reading comprehension usually comprises several sentences, leading to performance degradation caused by context length. Therefore, we model a hierarchical structure using sentence encoding, in which as the context becomes longer, the performance degrades. Experimental results show that the proposed VS³-NET model has an exact-match score of 76.8% and an F1 score of 84.5% on the SQuAD test set.

• Korean Journal of Radiology
• /
• 제23권10호
• /
• pp.949-958
• /
• 2022

Objective: To investigate the feasibility of using a deep learning-based analysis of auscultation data to predict significant stenosis of arteriovenous fistulas (AVF) in patients undergoing hemodialysis requiring percutaneous transluminal angioplasty (PTA). Materials and Methods: Forty patients (24 male and 16 female; median age, 62.5 years) with dysfunctional native AVF were prospectively recruited. Digital sounds from the AVF shunt were recorded using a wireless electronic stethoscope before (pre-PTA) and after PTA (post-PTA), and the audio files were subsequently converted to mel spectrograms, which were used to construct various deep convolutional neural network (DCNN) models (DenseNet201, EfficientNetB5, and ResNet50). The performance of these models for diagnosing ≥ 50% AVF stenosis was assessed and compared. The ground truth for the presence of ≥ 50% AVF stenosis was obtained using digital subtraction angiography. Gradient-weighted class activation mapping (Grad-CAM) was used to produce visual explanations for DCNN model decisions. Results: Eighty audio files were obtained from the 40 recruited patients and pooled for the study. Mel spectrograms of "pre-PTA" shunt sounds showed patterns corresponding to abnormal high-pitched bruits with systolic accentuation observed in patients with stenotic AVF. The ResNet50 and EfficientNetB5 models yielded an area under the receiver operating characteristic curve of 0.99 and 0.98, respectively, at optimized epochs for predicting ≥ 50% AVF stenosis. However, Grad-CAM heatmaps revealed that only ResNet50 highlighted areas relevant to AVF stenosis in the mel spectrogram. Conclusion: Mel spectrogram-based DCNN models, particularly ResNet50, successfully predicted the presence of significant AVF stenosis requiring PTA in this feasibility study and may potentially be used in AVF surveillance.

• 대한기계학회논문집A
• /
• 제22권1호
• /
• pp.90-101
• /
• 1998

This paper presents a scheduling method that uses Petri net modeling and heuristic search to handle the tool setup. In manufacturing systems, a tool is attached to a particular machine to process a particular operation. The activity to attach a tool to a particular machine and detach the tool from the machine requires time. The processing time of operations varies according to the attached tool and the machine used. The method proposed in this paper uses Petri net to model these characteristics and applies a search algorithm to the reachability graph of the Petri net model to generate an optimal or near-optimal schedule. New heuristic functions are developed for efficient search. The experimental results that show the effectiveness of the proposed method are presented.

• 제어로봇시스템학회논문지
• /
• 제8권8호
• /
• pp.706-712
• /
• 2002

In general, a deadlock in flexible manufacturing systems (FMSs) is caused by a resource limitation and the diversity of routings. However, the deadlock of industrial controllers such as programmable logic controllers (PLCs) can occur from different causes compared with those in general FMSs. The deadlock of PLCs is usually caused by an error signal between PLCs and manufacturing systems. In this paper, we propose a deadlock detection and recovery (DDR) algorithm to resolve the deadlock problem of PLCs at design stage. This paper employs the MAPN (modified automation Petri net), MTPL (modified token passing logic), and ECC (efficient code conversion) algorithm to model manufacturing systems and to convert a Petri net model into a desired LD (ladder diagram). Finally, an example of manufacturing systems is provided to illustrate the proposed DDR algorithm.